In the AK4499EXEQ data sheet, they have a recommended mono external circuit summing the two hot and cold lines together after a passive LPF stage. I am using an active LPF through an FDA and need to sum the signals before the LPF. What is the ideal way to sum two hot and two cold signals? The data sheet is unfortunately very vague when it comes to mono descriptions.
Okay, I'll the first stab at responding here.
IMHO, AKM likely suggested a particular approach to output stage design because they they think it tends to sound better to humans.
Why would passive filtering before summing be a good idea? For one thing, differential summing amps and or active balanced filters relay on both common mode and differential mode opamp input operation. The potential issue is that there is still some remaining RF in the dac analog output signals even after passing through the I/V converter stages. I have measured square wave clock signals buried in the RF shaped noise coming out of an original AK4499 dac after the I/V stage. It just so happens that OPA1612/1611 are pretty tolerant of the RF when used for I/V in inverting mode (where opamps historically tend to perform best).
Having said the above, some of it is opinion from personal experience, and some from private communications with professional dac designers. We independently came to believe that some extra passive filtering after the I/V stage was a good thing.
IMHO, AKM likely suggested a particular approach to output stage design because they they think it tends to sound better to humans.
Why would passive filtering before summing be a good idea? For one thing, differential summing amps and or active balanced filters relay on both common mode and differential mode opamp input operation. The potential issue is that there is still some remaining RF in the dac analog output signals even after passing through the I/V converter stages. I have measured square wave clock signals buried in the RF shaped noise coming out of an original AK4499 dac after the I/V stage. It just so happens that OPA1612/1611 are pretty tolerant of the RF when used for I/V in inverting mode (where opamps historically tend to perform best).
Having said the above, some of it is opinion from personal experience, and some from private communications with professional dac designers. We independently came to believe that some extra passive filtering after the I/V stage was a good thing.
Actually AK4499 summing same phases outputs in mono mode. following the LPF to XLR output.
But making balanced to Single ended line needs additional circuit also explained in the data-sheet.
And it has LPF too in the input Fo=112KHz so the passive LPF after IV conversion before XLR out probably is higher Fo than 112KHz to make sense of...
.
I can reccomend to move down a bit Fo for booth filters for instace for input to SE filter at 80KHz and for XLR respectivly higher Fo.
.
And there are some explanation of what happening inside about Iouts and other complementary pin with feedback formed inside the chip?
.
Note that the Virtual ground Voff is 2.5V not 0V. And that C is infact the bypassing C on Rfeedback.
.
BTW
There is a mistake in drawing by AKM in the first schematic... Last from top OP amp is not correctly connected. For the reference note second OP from top, 4th OP should be configured the same way.
.
cheers 🙂
But making balanced to Single ended line needs additional circuit also explained in the data-sheet.
And it has LPF too in the input Fo=112KHz so the passive LPF after IV conversion before XLR out probably is higher Fo than 112KHz to make sense of...
.
I can reccomend to move down a bit Fo for booth filters for instace for input to SE filter at 80KHz and for XLR respectivly higher Fo.
.
And there are some explanation of what happening inside about Iouts and other complementary pin with feedback formed inside the chip?
.
Note that the Virtual ground Voff is 2.5V not 0V. And that C is infact the bypassing C on Rfeedback.
.
BTW
There is a mistake in drawing by AKM in the first schematic... Last from top OP amp is not correctly connected. For the reference note second OP from top, 4th OP should be configured the same way.
.
cheers
Yes, feedback RC must be connected to OP's output, not the power line.
BTW, I think this is "copy-paste error", because the same mistake was also in old 4-channel 4499EQ, which is not manufactured last ~5 years, after the fire.
(And IMHO the new 4499EXEQ is much better, the old 4499eq was worse than the 4497, and even worse than the 4493!)
Alex.
Attachments
Last edited:
Yes
I checked with spice:
Output is 13Vp-p with AKM OP output.
So the Vrms value that they gave is correct.
But insane value of output for line device... What the preamplifier should do?
.
This value is for half current output (72.8mAp-p) 36.4mAp-p per one output
The Rfb is 360 ohms, so (0.0728mAp-p / 2) x 360 ohm = 13.1 Vp-p or 4.646 Vrms as in the datasheet...
Just to remind that some value of 2Vrms is common for DACs, and maybe it is also too much...
That was proven with spice.
So the value in datasheet of 72.7mAp-p is given as total value for 2 channels.
That was not clear from the datasheet... 🙁
.
The specified Fo for filter nets are close to one that stated and I think that they are correct.
I will post the transfer in db.
.
Other thing is that it could be the issue maybe for some OP amps to make IV with that high Io and lower Rfb?
Even for just one pair of Io
.
Another thng is maybe most important but it is not clear too.
Is there some permanent offset voltage on the current outputs?
(The DC correction voltage Voff = 2.5V in the schematic.)
I checked with spice:
Output is 13Vp-p with AKM OP output.
So the Vrms value that they gave is correct.
But insane value of output for line device... What the preamplifier should do?
.
This value is for half current output (72.8mAp-p) 36.4mAp-p per one output
The Rfb is 360 ohms, so (0.0728mAp-p / 2) x 360 ohm = 13.1 Vp-p or 4.646 Vrms as in the datasheet...
Just to remind that some value of 2Vrms is common for DACs, and maybe it is also too much...
That was proven with spice.
So the value in datasheet of 72.7mAp-p is given as total value for 2 channels.
That was not clear from the datasheet... 🙁
.
The specified Fo for filter nets are close to one that stated and I think that they are correct.
I will post the transfer in db.
.
Other thing is that it could be the issue maybe for some OP amps to make IV with that high Io and lower Rfb?
Even for just one pair of Io
.
Another thng is maybe most important but it is not clear too.
Is there some permanent offset voltage on the current outputs?
(The DC correction voltage Voff = 2.5V in the schematic.)
Common, but not standard (because there is no standard).
Most of my DC have the ootput from 2 to 3Vrms.
OP with enough output current or OP+2 BJT transistors may solve this problem.
Alex.
One reason for this is that DR (or SNR) is better with high output. It may be difficult to reach specified DR/SNR with 2Vrms output.
Current outputs are biased with VCOML/R. So by adjusting VCOML/R it is possible to adjust DC of I/V output. But this also has an impact on performance. Datasheet is silent about how this works internally.
Hi
Yes thanks for the info.
.
So the Iout-s have constant +Vdc value to the transimpedance OP stage of +2.5V
Because
VCOML/R = (VREFHL/R + VREFLL/R) / 2
and
VREFHL/R is typ. constant +5V
VREFLL/R is typ. constant +0V
Voff applied to the +in of OP by the datas have +2.5V = VCOML/R)
.
Some idea about spice circuit of Iouts?
The only confirmed info that matches with AKM data is that 36.4 mAp-p of each literary Io is true...
.
Additional info from PDF is that every pair of Iout and OPin have 30 ohm resistor internally
"IOUTRN Connected to OPINRN (30 Ω, typ.)"
.
Hi, sorry for not getting back to you. In the datasheet they parallel the two dacs in their mono mode external circuit. I intend to run this DAC in mono mode and I take it leaving one of the dacs floating is bad practise. Although the documentation as I said is very vague on this. They have also not responded to the two emails I sent them a few months back regarding this.
I know what is in the datasheet. My question was - why YOU want to use parallel (mono) mode? For what?
Alex.
Alex.
I guess what Altor is asking is what do you intend to gain with dual mono setup?
Regarding your original question what's wrong with using the datasheet circuit without the LPFs? So just parallel I/V outputs via 100 ohms resistors.
Regarding your original question what's wrong with using the datasheet circuit without the LPFs? So just parallel I/V outputs via 100 ohms resistors.